Systems and Methods for Hands-free Voice Control and Voice Search

• US 20110166855A1
• Filed: 07/06/2010
• Published: 07/07/2011
• Est. Priority Date: 07/06/2009
• Status: Active Grant

• Current Assignee: Sensory Incorporated
• Original Assignee: Sensory Incorporated
• Inventors: Shaw, Jonathan, Vermeulen, Pieter J., Mozer, Todd F.
• Granted Patent: US 8,700,399 B2
• Time in Patent Office: Days
• US Class Current: 704/231
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/081 Search algorithms, e.g. Baum-Welch or Viterbi
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

In one embodiment the present invention includes a method comprising receiving an acoustic input signal and processing the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound. Different acoustic recognition processes start processing different segments of the acoustic input signal at different time points in the acoustic input signal. In one embodiment, initial states in the recognition processes may be configured on each time step.

Claims

1. A method comprising:

receiving an acoustic input signal; and

processing the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound,

wherein different acoustic recognition processes start processing different segments of the acoustic input signal at different time points in the acoustic input signal.

2. The method of claim 1 wherein each acoustic recognition process generates a score, and wherein said word or phrase is recognized if at least one of said acoustic recognition processes has a score greater than a threshold when said at least one of said acoustic recognition processes is in a final state.

3. The method of claim 1 wherein the acoustic input signal is processed across a plurality of time steps, and wherein different acoustic recognition processes are started at successive time steps.

4. The method of claim 1 wherein each acoustic recognition process comprises a plurality of states corresponding to acoustic units, each state having an associated score.

5. The method of claim 4 wherein each acoustic recognition process comprises a state corresponding to an acoustic model for garbage.

6. The method of claim 4 wherein the plurality of states include one or more initial states, and wherein a score of the one or more initial states is reset on each frame if the score is less than a threshold.

7. The method of claim 6 wherein the score is reset to a predetermined value on each frame if the score is less than a threshold before calculating a current score for the initial state based on a received acoustic unit for a current frame.

8. The method of claim 7 wherein said predetermined value is a constant.

9. The method of claim 7 wherein different initial states are reset to different predetermined values.

10. The method of claim 4 wherein the score comprises a first component indicating a probability of a match between a particular state and a received acoustic unit of the acoustic input signal, a second component comprising a highest score of a predecessor state, and a third component comprising an offset.

11. The method of claim 4 wherein the score comprises an offset to increase the score of each state so that scores of different durations are comparable.

12. The method of claim 11 wherein the offset is a constant greater than one.

13. The method of claim 11 wherein different states have different associated offset values.

14. The method of claim 4 wherein the plurality of states include one or more final states, and wherein a result is generated when a score of a final state increases above a threshold.

15. The method of claim 1 wherein the acoustic recognition processes operate on multiple words in parallel.

16. A method comprising:

receiving, in a processor, an acoustic input signal; and

processing the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound, the plurality of acoustic recognition processes comprising a Viterbi search of a plurality of states corresponding to acoustic units of an acoustic model, the plurality of states including initial states and final states of the acoustic model, wherein

initial states are reset on each time steps if a score for the initial state on a previous time step is below a threshold,

said score is calculated for a plurality of said states on each time step, said score comprising an offset to increase the score of each state so that scores of different durations are comparable, and

a result is generated when a score of a final state increases above a threshold.

17. A computer readable storage medium embodying a computer program for performing a method, said method comprising:

receiving an acoustic input signal; and

processing the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound,

wherein different acoustic recognition processes start processing different segments of the acoustic input signal at different time points in the acoustic input signal.

18. The computer readable storage medium of claim 17 wherein each acoustic recognition process comprises a plurality of states corresponding to acoustic units, each state having an associated score.

19. The computer readable storage medium of claim 18 wherein the plurality of states include one or more initial states, and wherein a score of the one or more initial states is reset on each frame if the score is less than a threshold.

20. The computer readable storage medium of claim 18 wherein the score comprises an offset to increase the score of each state so that scores of different durations are comparable.

21. The computer readable storage medium of claim 18 wherein the plurality of states include one or more final states, and wherein a result is generated when a score of a final state increases above a threshold.

22. A computer system comprising:

one or more processors, each processor configured to;

receive an acoustic input signal; and

process the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound,

wherein different acoustic recognition processes start processing different segments of the acoustic input signal at different time points in the acoustic input signal.